cv2.error: OpenCV(4.1.0) C:projectsopencv-pythonopencvmodulesimgprocsrcse

2021-01-18 11:43:52 浏览数 (1)

参考链接: Python Opencv 基础3 : resize 调整图像大小

问题描述 

在运用分水岭分割的时候,发现该常见错误 

kernel = np.ones((3, 3), np.uint8)

sure_bg = cv.dilate(waterimg, kernel, iterations=3)# 确定背景

dist = cv.distanceTransform(waterimg, cv.DIST_L2, 3)# 距离变换

ret, sure_fg = cv.threshold(dist, dist.max()*0.6, 255, 0)# 确定前景

surface_fg = np.uint8(sure_fg)

unknown = cv.subtract(sure_bg, surface_fg)

ret, markers = cv.connectedComponents(surface_fg)

# markers = np.uint8(markers)

markers = markers 1

markers[unknown == 255] = 0

waterimg[markers == -1] = 255

markers = cv.watershed(gray, markers)

错误名称: 

   cv2.error: OpenCV(4.1.0) C:projectsopencv-pythonopencvmodulesimgprocsrcsegmentation.cpp:161: error: (-215:Assertion failed) src.type() == CV_8UC3 && dst.type() == CV_32SC1 in function 'cv::watershed'

意思是该函数正在尝试将8通道的转为32通道 

解决方法 

markers = cv.watershed(img, markers)

该函数中img必须为三通道,即不能为灰度图或二值图像,可以用cvtcolor将gray2bgr,这样就不会出错了。

0 人点赞